General Freshness Recognition Method Based on Electronic Nose and Improved Unsupervised Signature Projection Algorithm.

The odor information of living hairy crab was collected by self-made electronic nose system. The dimension of multi-dimensional characteristic response of hairy crab samples was reduced by popular learning algorithm, and the low-dimensional feature vector of samples was extracted. Then, the freshness of hairy crab was identified by back-propagation neural network, and compared with the physical and chemical index volatile basic nitrogen. The results show that the accuracy of the algorithm can reach 98.1%, and the results based on the electronic nose technology and the physical and chemical indicators are basically the same. Therefore, the non-destructive identification method based on electronic nose technology is feasible.